In an agent-written codebase the agent-written tests are equally untrustworthy; manually using the product is the only reliable measure of whether it works.

Drawn from what Mario Zechner said

On the topicWhat are tests for, and how much testing is enough?under our view: Tests an agent wrote cannot be trusted on their own, so a person still has to check On the topicWhat do coding agents change about software work?under our view: The person who checks the work is a safety property, and removing them is the risk

What Mario Zechner actually said

Do you hold this korrent?Do you also believe this?

Cannot both be true

Tests written by a model in the same context as the change are what find the bugs; the automated tests left behind are the lesser product.

Steinberger has the model write tests in the same context as the change because that is what finds the bugs. Zechner says the tests an agent writes for its own code are as untrustworthy as the code, and that manually using the product is the only reliable measure left. They cannot both be right about whether a model can check its own work.
In tension with

In most domains agents are now better than humans at finding bugs.

Hansson says agents are now better than humans at finding bugs in most domains. Zechner says the tests an agent writes for agent-written code are as untrustworthy as the code, and that manually using the product is the only reliable measure left. Better at finding bugs and unable to certify its own work are compatible, and the gap is exactly where a team decides how much to review.

Shared subject first, then similar wording. They may agree or disagree.